Autonomic Reactivity and EEG-defined Clusters Reveal Distinct Chronic Pain Profiles

OBJECTIVES: This study examined whether integrating stress-related heart rate variability (HRV) with resting-state electroencephalography (EEG) could characterize physiological subgroups in Chronic pain after spine surgery (CPSS). METHODS: A total of 177 patients completed pain- and emotion-related questionnaires, a 3-minute resting-state EEG recording, and HRV assessments during baseline, mental arithmetic stress, and recovery. Participants were grouped by the direction of stress-related low-frequency/high-frequency (LF/HF) ratio change. Principal component analysis and unsupervised clustering were applied to 437 EEG features encompassing absolute power, cerebral asymmetry, and theta/beta ratio (TBR). After multivariate outlier detection, 168 participants underwent clustering. A 2 × 2 factorial ANOVA examined HRV-group and EEG-cluster effects on pain scores. RESULTS: LF/HF -decreased group (n = 56) reported higher depressive symptom scores than the increase group (n = 121, P = 0.045), without significant pain differences. Among 437 EEG features, PC1 reflected beta activity (34.39% variance), PC2 frontal/frontotemporal asymmetry (14.52%), and PC3 TBR features (9.55%). Two EEG clusters (n = 88 and n = 80) showed no significant pain- or emotion-related differences. Factorial analysis revealed nonsignificant HRV (P = .056) and EEG main effects (P = .107), but a nominally significant interaction [P = .030, partial η² = .028 (HC3-robust P = .044)]. Within EEG Cluster 0, the LF/HF-decrease group reported higher pain scores than the LF/HF-increase group (70.9 vs. 42.3; Tukey-adjusted P = .019). DISCUSSION: Integrating HRV-EEG profiles may help characterize CPSS heterogeneity. These exploratory findings, without study-wide multiplicity adjustment, require independent validation before clinical application.